I don't see why 1) is unlikely - people that work with bats have antibodies to bat (corona)viruses already, and we know that even in a vacuum people often have very mild symptoms or no symptoms at all. It's even more likely since the virus was probably not adapted to human hosts initially.
2) is also likely, for some viruses it took years and years, sometimes even more than a decade to find the actual intermediate animal.
4) is unlikely because further analysis cannot produce a likely scenario. If the virus was from an animal source known to the lab, we would know already, and if it was due to a gain-of-function experiment, it would be quite unlikely for the virus to take so much time to adapt to humans (it still hasn't fully done so), and there is still a lot of function to be gained. Besides, there is no obvious marker for genetic engineering (the furin cleavage sites are perfectly well described by both 1 and 2), and the fact the virus does not seem to be at a local optima yet indicates that it's probably not the result of engineering by repeated selection.
2) is also likely, for some viruses it took years and years, sometimes even more than a decade to find the actual intermediate animal.
4) is unlikely because further analysis cannot produce a likely scenario. If the virus was from an animal source known to the lab, we would know already, and if it was due to a gain-of-function experiment, it would be quite unlikely for the virus to take so much time to adapt to humans (it still hasn't fully done so), and there is still a lot of function to be gained. Besides, there is no obvious marker for genetic engineering (the furin cleavage sites are perfectly well described by both 1 and 2), and the fact the virus does not seem to be at a local optima yet indicates that it's probably not the result of engineering by repeated selection.